Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
RESEARCHPAPER
How do climate and dispersal traits limitranges of tree species along latitudinaland elevational gradients?Andrew Siefert*, Mark R. Lesser† and Jason D. Fridley
Department of Biology, Syracuse University,
107 College Place, Syracuse, NY 13244, USA
ABSTRACT
Aim We compared the upper range limits of tree species along latitudinal andelevational gradients to assess the influence of specific climatic factors – wintertemperature, summer temperature and growing season length – in determiningspecies range limits. We analysed the degree and direction of climatic mismatchesbetween latitudinal and elevational limits to determine whether mismatches couldbe explained by species dispersal traits.
Location Eastern North America and the Great Smoky Mountains (Tennesseeand North Carolina), USA.
Methods We determined the climatic limits for each of 28 common tree speciesalong a latitudinal gradient in eastern North America and across an elevationalgradient in the Great Smoky Mountains. We calculated the degree of climaticmismatch between species limits along the two gradients and tested for relation-ships with species dispersal traits (seed mass, maximum height, seed and pollendispersal mode).
Results We found strong positive relationships between species latitudinal andelevational climatic limits. Winter temperatures were much lower at speciespoleward limits compared with their upper elevational range limits, but there wereclose matches between limits for variables related to summer temperature andgrowing season length. The degree of climatic mismatch was influenced by speciesdispersal traits; species with unassisted seed dispersal and low maximum heighttended to have lower latitudinal than elevational limits.
Main conclusions Our results suggest that low summer temperatures and ashort growing season length limit species distributions along both latitudinal andelevational gradients, whereas winter temperature does not play a critical role. Thefailure of some species with poor dispersal ability to grow as far poleward asexpected based on their elevational limits supports the hypothesis that historicaldispersal constraints may limit species distributions along the latitudinal gradient.
KeywordsDispersal limitation, distribution, eastern North America, environmentalgradients, gene flow, Great Smoky Mountains, growing season, migration lag,range limit.
*Correspondence: Andrew Siefert, Departmentof Evolution and Ecology, University ofCalifornia, One Shields Avenue, Davis, CA95616, USA.E-mail: [email protected]†Present address: Department of Environmentaland Forest Biology, SUNY-ESF, One ForestryDrive, Syracuse, NY 13210, USA.
INTRODUCTION
The turnover of plant and animal species along environmental
gradients is one of the oldest themes in ecology (Schimper, 1898;
Whittaker, 1956; von Humboldt, 2009). Yet understanding the
controls on species distributions that produce this turnover
remains a central question in ecology, with important implica-
tions for predicting species range shifts in response to ongoing
and predicted climate change (Jump et al., 2009; Chen et al.,
2011).
Climate has long been recognized as a primary factor control-
ling the distributions of tree species, especially at their upper
bs_bs_banner
Global Ecology and Biogeography, (Global Ecol. Biogeogr.) (2015) 24, 581–593
© 2015 John Wiley & Sons Ltd DOI: 10.1111/geb.12287http://wileyonlinelibrary.com/journal/geb 581
latitudinal and elevational limits (Tranquillini, 1979; Huntley
et al., 1989; Randin et al., 2013). Several climate-based limita-
tions have been proposed, including low winter temperatures
(Sakai & Weiser, 1973), low growing season temperatures
(Pigott & Huntley, 1981; Cogbill & White, 1991; Mellert et al.,
2011) and a shortened growing season length (Morin et al.,
2007; Normand et al., 2009). However, in many cases it remains
unclear which climate variables, and interactions between them,
are responsible for setting species range limits.
A comparison of climatic thresholds along latitudinal and
elevational gradients may provide insights into the factors con-
trolling species range limits (Kollas et al., 2014). A close corre-
spondence between values of a given climatic factor at the
latitudinal and elevational limits of a species would suggest that
that factor plays an important role in determining those limits.
In contrast, mismatch in a given factor between latitudinal and
elevational limits would suggest it is not the primary limit on
species distributions along one or both gradients. Recent studies
comparing cold limits of plant species along elevational gradi-
ents in the Alps and latitudinal gradients in Europe have found
positive relationships and close correspondences for most
species (Halbritter et al., 2013; Randin et al., 2013), providing
evidence of thermal constraints. Importantly, climatic factors
differ in their relative rates of change along latitudinal and
elevational gradients. In particular, sites at high elevations tend
to have warmer winter minimum temperatures and longer
growing seasons than high-latitude lowland sites with similar
growing season temperatures (Billings, 1973). This partial
decoupling of climatic changes along the two gradients allows
assessment of which climatic variables match most closely
between species latitudinal and elevational ranges limits, and
thus are most likely to control those limits. Climatic compari-
sons of high-elevation and high-latitude tree lines have generally
found the closest correspondence in growing season tempera-
ture (Cogbill & White, 1991; Jobbágy & Jackson, 2000; Körner &
Paulsen, 2004), but whether this pattern extends to individual
tree species is not known.
Although climate is thought to be the primary driver of high-
latitude and high-elevation range limits of tree species, the influ-
ence of other factors may generate climatic mismatches (Randin
et al., 2013). A host of potentially limiting abiotic factors change
differently with increasing elevation and latitude, and most of
these differences are expected to create more stressful conditions
for plants at high elevations. In particular, high-elevation sites
are characterized by increased steepness, exposure, cloud cover,
wind speed, ice deposition, UV radiation and diurnal tempera-
ture fluctuations and reduced soil depth, nutrient availability,
CO2 partial pressure and summer photoperiod – all factors that
may keep tree species below their expected thermal limits
(Billings, 1973). For example, Cogbill & White (1991) found
that upper limits of deciduous forest and tree lines in mountain
ranges in eastern North America were held below the expected
climate-determined elevations due to the effects of exposure. If
non-climatic abiotic stresses limit tree species distributions
more strongly along elevational than latitudinal gradients,
species elevational limits should be lower than their latitudinal
climatic limits (Fig. 1).
Another important difference between latitudinal and
elevational gradients is that elevational gradients are much
more spatially compressed, such that equivalent changes in
temperature occur over much shorter distances (Jump et al.,
2009). As a result, there is potential for more efficient seed and
pollen dispersal between contrasting climates in mountainous
regions (Sundblad & Andersson, 1995). This difference in dis-
persal may generate climatic mismatches between latitudinal
and elevational range limits in several ways. First, evidence
suggests that the expansion of tree species in Europe and
North America following glacial retreat was dispersal limited
(Davis et al., 1986; Svenning et al., 2008), such that species
poleward range limits lag behind their potential climatic limits
(Svenning & Skov, 2004). In contrast, dispersal should be
much less limiting of species ranges along spatially compressed
elevational gradients (Jump et al., 2009). If this is the case, tree
species should reach colder temperatures along elevational
than latitudinal gradients (Randin et al., 2013), and this mis-
match should be greater for species with poorer seed dispersal
abilities (Fig. 1). Second, depending on the strength of selec-
tion, gene flow from core (i.e. low elevation) to peripheral (i.e.
high elevation) populations may prevent local adaptation at
range margins (García-Ramos & Kirkpatrick, 1997; Savolainen
et al., 2007), potentially limiting the cold tolerance and upper
elevational range of tree species (Jump et al., 2009). If this is
Figure 1 Illustration of the hypothesizedcauses of climatic mismatch between theupper latitudinal and elevational limits oftree species. Contour lines represent thepotential upper limit of a species along theelevational or latitudinal gradient based onits climatic tolerances. Arrows representexpected shifts in species limits above orbelow their potential climatic limit due tohypothesized mechanisms.
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd582
the case, species should reach colder temperatures along lati-
tudinal than elevational gradients, and the mismatch should be
greater for species with greater seed or pollen dispersal abilities
(Fig. 1). In contrast, efficient dispersal along elevational
gradients may allow populations to exceed their long-term
climatic limits through source–sink dynamics (Pulliam, 1988;
Halbritter et al., 2013), and if this is the case, species should
reach colder temperatures along elevational than latitudinal
gradients, and the mismatch should be greater for species with
greater seed dispersal abilities (Fig. 1).
Comparison of observed patterns of climatic mismatch
between species elevational and latitudinal limits with theoreti-
cal predictions has the potential to provide novel insights into
the roles of climate and dispersal in shaping species distribu-
tions. In this study, we examined the range limits of 28 common
tree species along a latitudinal gradient in eastern North
America and an elevational gradient in the Great Smoky Moun-
tains of the southern Appalachians. Our specific objectives were:
(1) to determine the relationship and degree of matching
between species latitudinal and elevational climatic limits; (2) to
assess the influence of specific climatic factors – winter tempera-
ture, summer temperature and growing season length – in deter-
mining these limits; and (3) to determine whether mismatches
between latitudinal and elevational limits could be explained by
species dispersal traits.
METHODS
Study area
The study was conducted across a latitudinal gradient in eastern
North America (25–50° N, 60–95° W). This gradient of over
1500 km encompasses the entire range of many eastern North
American tree species, transitioning from southern pines and
evergreen broadleaved species to mid-latitude broadleaved
deciduous species and boreal evergreen conifers in the north.
For comparison with the latitudinal gradient we used an
elevational gradient in the Great Smoky Mountains National
Park located in the southern Appalachians of Tennessee and
North Carolina, USA (35°26′–35°47′ N, 83°3′–84°0′ W; Fig. 2).
The Great Smoky Mountains represent the southern extent of
high peaks in eastern North America, with elevations ranging
from 224 to 2024 m a.s.l. This range of elevations occurs over
relatively short distances (about 15 km), creating a short steep
gradient along which the turnover of tree species mirrors that of
the latitudinal gradient across eastern North America (Shanks,
1954). The elevational gradient in the Great Smoky Mountains
and the latitudinal gradient across eastern North America cover
a similar range of summer temperature (mean July temperature
ranges 14.2–26.7 °C and 13.9–28.5 °C, respectively) and
growing season length (annual frost-free period 207–330 days
and 142–360 days, respectively), but the range of winter tem-
peratures is much smaller in the Great Smoky Mountains than
in eastern North America (mean January temperature ranges
0.0–4.7 °C and −10.0 to 19.9 °C, respectively; Shanks, 1954;
Busing et al., 2005; Fridley, 2009).
Species distributions
Data on tree species distributions in the Great Smoky Mountains
National Park originated from a systematic vegetation survey of
the park conducted from 1935 to 1938 (MacKenzie & White,
1998). The survey consisted of 1378 rectangular 0.08-ha plots
located randomly within the park. This remains the only study to
date that has sampled the entire park, encompassing the full
elevational gradient. In each plot, woody plants were identified to
species and classified as trees (> 10 cm diameter at breast height)
or shrubs. For our analysis, we included only species that
occurred as trees in at least 50 plots and had an upper elevational
limit within the park, which we defined as having a maximum
elevation below the 95th percentile of plots in the dataset. The
final species list included 28 species, representing many of the
most common hardwood and conifer species in the Great Smoky
Mountains and eastern North America (Table 1).
For tree species distributions in eastern North America we
used digitized versions of Little’s (1971) tree species range maps
Figure 2 Map of the study area showing (a) a digital elevationmodel of eastern North America with the location of the GreatSmoky Mountains National Park (GSMNP) and (b) enlargementof the GSMNP displaying elevational changes of > 1500 m withina spatial extent of < 10 km.
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 583
Tab
le1
Clim
atic
limit
s(2
.5%
quan
tile
s)of
28st
udy
spec
ies
alon
gla
titu
din
algr
adie
nt
inea
ster
nN
orth
Am
eric
a(E
NA
)an
del
evat
ion
algr
adie
nt
inth
eG
reat
Smok
yM
oun
tain
s(G
SM)
base
don
Jan
uar
ym
ean
tem
pera
ture
,Ju
lym
ean
tem
pera
ture
and
fros
tfr
eeda
ys(F
FD).
Th
em
ism
atch
betw
een
the
lati
tudi
nal
and
elev
atio
nal
limit
isgi
ven
for
each
clim
ate
vari
able
.Pos
itiv
em
ism
atch
esin
dica
teel
evat
ion
allim
its
are
cold
er(o
rh
ave
few
erFF
D)
than
lati
tudi
nal
limit
s.N
egat
ive
mis
mat
ches
indi
cate
elev
atio
nal
limit
sar
ew
arm
er(o
rh
ave
mor
eFF
D)
than
lati
tudi
nal
limit
s.Sp
ecie
sar
eso
rted
from
the
hig
hes
tto
the
low
est
elev
atio
nal
limit
.
Spec
ies
Jan
uar
ym
ean
tem
p.(°
C)
July
mea
nte
mp.
(°C
)FF
D(d
ays
year
−1)
Lat
itu
din
al
limit
Ele
vati
onal
limit
Mis
mat
ch
Lat
itu
din
al
limit
Ele
vati
onal
limit
Mis
mat
ch
Lat
itu
din
al
limit
Ele
vati
onal
limit
Mis
mat
ch
Bet
ula
lent
a−7
.59
−1.3
7−6
.22
18.6
516
.56
2.09
130.
0512
0.24
9.81
Ace
rsa
ccha
rum
−15.
95−1
.29
−14.
6616
.37
16.6
9−0
.32
97.8
912
0.73
−22.
84
Aes
culu
sfla
va−2
.17
−1.2
8−0
.88
20.1
616
.70
3.46
154.
8512
0.94
33.9
1
Ace
rpe
nsyl
vani
cum
−14.
24−1
.26
−12.
9816
.32
16.7
4−0
.42
109.
4112
1.38
−11.
98
Pru
nus
sero
tina
−11.
60−1
.25
−10.
3618
.25
16.7
71.
4812
4.53
121.
792.
73
Tsug
aca
nade
nsis
−13.
28−1
.15
−12.
1317
.13
16.9
20.
2111
3.87
122.
43−8
.57
Hal
esia
tetr
apte
ra−0
.47
−0.9
70.
5020
.25
17.2
33.
0215
5.90
125.
4030
.50
Ace
rru
brum
−15.
37−0
.91
−14.
4516
.72
17.3
2−0
.60
106.
9712
6.83
−19.
86
Que
rcus
rubr
a−1
5.28
−0.9
1−1
4.36
16.9
317
.32
−0.4
010
8.05
126.
83−1
8.78
Rob
inia
pseu
doac
acia
−3.6
9−0
.76
−2.9
319
.82
17.5
72.
2515
3.09
127.
1725
.93
Frax
inus
amer
ican
a−1
2.97
−0.7
5−1
2.22
17.4
217
.59
−0.1
711
8.88
127.
21−8
.33
Tili
aam
eric
ana
−14.
94−0
.62
−14.
3217
.48
17.8
0−0
.33
110.
5213
3.70
−23.
18
Mag
nolia
fras
eri
−4.1
8−0
.56
−3.6
218
.52
17.9
10.
6114
7.24
133.
9913
.25
Que
rcus
alba
−9.7
8−0
.52
−9.2
619
.26
17.9
81.
2813
9.26
137.
371.
89
Mag
nolia
acum
inat
a−5
.08
−0.4
1−4
.67
19.3
118
.17
1.15
139.
8913
9.40
0.49
Sass
afra
sal
bidu
m−5
.21
−0.4
0−4
.82
20.3
218
.19
2.13
154.
3913
9.40
14.9
9
Que
rcus
mon
tana
−5.7
3−0
.39
−5.3
319
.71
18.1
91.
5214
8.01
139.
408.
61
Que
rcus
velu
tina
−7.7
5−0
.32
−7.4
320
.10
18.3
11.
7915
1.44
139.
6411
.80
Oxy
dend
rum
arbo
reum
−1.7
6−0
.19
−1.5
720
.71
18.5
32.
1716
1.84
146.
2715
.57
Nys
sasy
lvat
ica
−5.2
7−0
.19
−5.0
820
.29
18.5
41.
7615
3.76
146.
317.
46
Car
yagl
abra
−5.2
5−0
.15
−5.1
020
.64
18.6
12.
0415
7.12
147.
0010
.13
Car
yaal
ba−4
.41
−0.0
3−4
.39
20.6
918
.80
1.89
157.
1714
9.53
7.63
Pin
usri
gida
−7.4
50.
01−7
.46
19.2
918
.86
0.43
141.
3214
9.64
−8.3
2
Cor
nus
flori
da−4
.92
0.04
−4.9
720
.56
18.9
21.
6415
5.44
149.
875.
57
Liri
oden
dron
tulip
ifer
a−5
.22
0.09
−5.3
120
.20
19.0
01.
2015
4.47
150.
603.
87
Que
rcus
cocc
inea
−5.6
20.
10−5
.72
19.8
819
.01
0.87
149.
0815
0.75
−1.6
7
Pin
usst
robu
s−1
6.90
0.65
−17.
5516
.38
19.9
4−3
.55
98.3
016
0.91
−62.
61
Pin
usvi
rgin
iana
−3.9
60.
66−4
.62
19.9
719
.95
0.02
154.
3016
0.93
−6.6
3
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd584
(http://esp.cr.usgs.gov/data/little). Although dated, these maps
remain the definitive source for the ranges of North American
tree species. We compared all results using these data with an
alternative data source of species distributions using plot data
collected between 2001 and 2006 from the United States Depart-
ment of Agriculture Forest Inventory and Analysis Program
(FIA; http://www.fia.fs.fed.us) and Canada’s National Forest
Inventory (NFI; https://nfi.nfis.org). Both approaches produced
similar results, but we only report results based on Little’s
species range maps because we believe these provide more accu-
rate estimates of species poleward range limits due to increas-
ingly sparse coverage of Canadian plot data with increasing
latitude. Sparse plot coverage at higher latitudes resulted in the
poleward range limits of some species being at lower latitudes
than limits provided by Little’s range maps.
Climate data
We extracted daily minimum and maximum temperatures from
1980 to 2010 for each 0.5° × 0.5° grid cell in eastern North
America from Daymet (http://daymet.ornl.gov). We obtained
similar daily temperature data from six weather stations in or
near Great Smoky Mountains National Park that spanned the
entire elevational gradient of the park (Fridley, 2010). Weather
station data were then used to calculate linear lapse rates of
temperature along the elevational gradient which enabled us to
calculate daily minimum and maximum temperatures for each
of the vegetation plots in the Great Smoky Mountains. While
these climate data are not from the same time period as the
species distribution data we used, temperatures in the Great
Smoky Mountains have remained relatively stable over the past
century (Fig. S1 in Supporting Information), meaning that
these estimates of species climatic distributions were accurate.
Further, Lesser & Fridley (unpublished data) found a lack of
warming in the Great Smoky Mountains that agrees with other
studies that have observed no significant warming across much
of the south-eastern United States (Portmann et al., 2009; Meehl
et al., 2012).
Using the daily temperature data, we derived climatic vari-
ables representing factors hypothesized to control species upper
latitudinal and elevational range limits: winter cold, growing
season length and growing season temperature. To quantify
winter cold, we calculated the mean daily temperature of the
coldest month (January) and the mean annual absolute
minimum temperature. We quantified the length of the growing
season using two metrics: (1) annual frost-free days (FFD), the
period between the last frost (daily minimum temperature
< 0 °C) of the spring and first frost of the autumn, and (2)
growing season length (GSL), the period beginning when daily
mean temperature exceeded 5 °C for five consecutive days and
ending when daily mean temperature fell below 5 °C for five
consecutive days (Kollas et al., 2014). We quantified growing
season temperature using two metrics: (1) mean daily tempera-
ture during the warmest month (July), and (2) growing season
temperature (GST), the mean daily temperature during the
growing season as defined above. We also calculated the number
of annual growing degree-days (GDD), which takes into account
both the length and temperature of the growing season, as the
cumulative annual sum of daily mean temperature minus a base
temperature of 10 °C.
For reasons of data availability, we used air temperature to
quantify potential thermal constraints on tree species, but pre-
vious work has found that soil temperature rather than air tem-
perature is critical for determining the position of the tree line
(Körner & Paulsen, 2004). However, root-zone and canopy tem-
peratures are highly correlated, particularly when integrated
over periods of a week or more (Körner & Paulsen, 2004), so we
believe that the use of soil temperature data would not have had
a strong impact on our findings.
Data analysis
For each species, we calculated climatic limits along the latitu-
dinal and elevational gradients as the 0%, 2.5% and 5%
quantiles of the distribution of climatic values of species occur-
rences in eastern North America and the Great Smoky Moun-
tains. Results of subsequent analyses were similar regardless of
which quantile was used, so we only report results based on the
2.5% quantiles.
We tested the relationship between species climatic limits
along latitudinal and elevational gradients using major axis
regression. We also calculated the mismatch between latitudinal
and elevational limits of each species for each climatic variable
as the difference between the latitudinal and elevational climatic
limit. Positive mismatches indicate that species limits are higher,
and thus colder, along the elevational gradient than the latitu-
dinal gradient. Conversely, negative mismatches indicate that
species limits are lower, and thus warmer, along the elevational
gradient than the latitudinal gradient. We tested whether the
average mismatch across all species for each climatic variable
was significant using paired t-tests. To visually assess mis-
matches, we superimposed species elevational climatic limits
onto the latitudinal gradient, and vice versa. For each species,
this allowed us to create contour lines indicating a species’
expected range limit along a given gradient based on its climatic
limits on the other gradient (Figs 3 & S2).
Next, we assessed whether the degree of mismatch could be
explained by species dispersal traits. We selected four traits
related to dispersal ability: maximum height, seed mass and seed
(wind, animal or unassisted) and pollen (wind or animal) dis-
persal mode. Seed dispersal distance is known to be positively
related to plant height and negatively related to seed mass
(Nathan et al., 2002; Thomson et al., 2011). Wind-dispersed
seed has been shown to generally disperse much farther than
animal-dispersed or unassisted seed (Wilson & Traveset, 2000),
although this trend is complicated by some animals (birds) gen-
erally having much greater dispersal distances than others such
as ants (Horvitz & Le Corff, 1993). Species were also categorized
as having wind- or animal-assisted pollen dispersal, with pollen
dispersal distances expected to be greater for wind-dispersed
species. Dispersal trait data were derived from Burns & Honkala
(1990) and the Seed Information Database (Royal Botanic
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 585
f) Pinus strobus
5728
500
1000
1500
2000
elev. (metres a.s.l.)
e) Liriodendron tulipifera
500
1000
1500
2000
elev. (metres a.s.l.)
2627
d) Quercus alba
3838
500
1000
1500
2000
elev. (metres a.s.l.)
b) Aesculus flava
500
1000
1500
2000
elev. (metres a.s.l.)
c) Acer rubrum
5322
500
1000
1500
2000
elev. (metres a.s.l.)
a) Betula lenta
500
1000
1500
2000
3257 elev. (metres a.s.l.)
Jan mean temp
July mean tempFFD
Figure 3 Distributions of six representative tree species, (a) Betula lenta, (b) Aesculus flava, (c) Acer rubrum, (d) Quercus alba, (e)Liriodendron tuliperifera, and (f) Pinus strobus, in eastern North America and along an elevational gradient in the Great Smoky Mountains(inset profile). Species distribution is shown with dark grey shading. Contour lines on maps show a species’ expected poleward range limitbased on its corresponding climatic limit along the elevational gradient. Horizontal lines on elevation profiles show a species’ expectedupper limit based on its corresponding latitudinal climatic limit in eastern North America. Climatic limits are shown for January meantemperature (dotted line), July mean temperature (solid line) and frost free days (dashed line).
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd586
Gardens Kew, 2014). We tested relationships between species
dispersal traits and elevational versus latitudinal climatic limit
mismatch using linear regressions.
Our approach assumes that dispersal ability is related to a
species’ standard seed dispersal mode. However, non-standard
dispersal vectors have been shown to contribute substantially to
long-distance dispersal (Higgens et al., 2003; Nathan et al.,
2008). For example, blue jays may have played an important role
in the post-glacial migration of fagaceous species (Quercus,
Fagus) otherwise considered to be poor dispersers (Johnson &
Webb, 1989). Long-distance dispersal events, however, represent
only a small proportion of dispersal events (Nathan, 2006). Our
conservative estimates of species range margins, based on
Little’s range maps (Little, 1971), do not account for outlying,
disjunct populations, which may be more greatly influenced by
long-distance and non-standard dispersal. Thus, species distri-
bution limits, as we have calculated them here, are more likely to
be dictated by local dispersal using standard mechanisms.
RESULTS
Climatic limits and mismatches
Overall, there was a positive relationship between species cli-
matic limits along latitudinal and elevational gradients, with
species that reached colder temperatures along a latitudinal gra-
dient also having colder elevational limits for each of the
included climate variables (Table 1, Fig. 4). The one significant
outlier was Pinus strobus (white pine) which had a much warmer
elevational limit than would be predicted by its latitudinal limit
(Table 1, Fig. 3f). At the opposite extreme (although not a sig-
nificant outlier), Aesculus flava (yellow buckeye) consistently
had a latitudinal limit that was warmer than predicted by the
elevational limit (Table 1, Fig. 3b).
Significant positive relationships between latitudinal and
elevational limits were found for all tested climate variables
(Table 2); however, because of high correlations between climate
variables we focus the remaining analysis on January mean tem-
perature, July mean temperature and FFD, as these had the
strongest relationships and lowest mismatch values among vari-
ables representing winter temperature, growing season tempera-
ture and growing season length, respectively. (Table 2).
Of these climatic variables, mismatches between latitudinal
and elevational range limits were greatest for January mean
temperature (Table 1), with species reaching much colder
winter temperatures along the latitudinal gradient than the
elevational gradient (Figs 3, 4a & S2). The average January
mean temperature at the poleward limit across study species
was −8 °C, while the average at the upper elevational limit was
−0.5 °C, resulting in an average mismatch across all species of
−7.6 °C (paired t-test, P < 0.001; Table 2). The mismatch
tended to be greater for species with more poleward and
higher limits (Fig. 4a).
Elevational and latitudinal limits of July mean temperature
and FFD both showed much closer correspondence. Species
limits for July mean temperature were strongly correlated, with
a major axis regression slope of 0.38 (0.52 with P. strobus
removed; Table 2). The average July mean temperature at the
poleward limit across study species was 19 °C, compared with
18 °C at the upper elevational limit (Table 1), resulting in an
average mismatch of < 1 °C (paired t-test, P = 0.006; Table 2).
There was a slight tendency for species to have colder elevational
than latitudinal limits, and these mismatches tended to be
greater for species with more equatorial latitudinal limits and
lower elevational limits (Fig. 4b).
Species limits for FFD were also strongly correlated, with the
slope from the major axis regression being 0.35 (0.52 with
P. strobus removed; Table 2). The average number of FFD at the
poleward limit was 137.4 days while the average at the upper
elevational limit was 137 days (Table 1), producing a non-
significant mismatch of only 0.4 days (paired t-test, P = 0.93;
Table 2). For FFD, species were more equally split on either side
of the 1:1 line, as opposed to the other variables where most
species fell on one side or the other (Fig. 4). Species with more
equatorial and lower limits tended to have a shorter growing
season (fewer FFD) at their elevational than their latitudinal
limit (Fig. 4c). Conversely, species that occurred farther
poleward and higher on the elevational gradient tended to have
a longer growing season (more FFD) at their elevational than
their latitudinal limit (Fig. 4c).
Effects of dispersal traits on climatic mismatch
Linear regression showed no significant relationship between
climatic mismatch and either seed mass or pollination mode
(Table S2). Dispersal mode and maximum height both showed
significant relationships with climatic mismatch (Fig. 5,
Table S2). For all climatic variables, the mismatch between lati-
tudinal and elevational limits became more negative (i.e. latitu-
dinal limits became colder relative to elevational limits) with
increasing height and with the transition from unassisted seed
dispersal to animal dispersal to wind dispersal (Fig. 5). Mis-
matches were negative across all dispersal groups and heights for
January mean temperature, indicating that the latitudinal limits
were colder than elevational limits. However, species with unas-
sisted seed dispersal (relatively poor dispersers) had the smallest
levels of mismatch, whereas wind-dispersed species (relatively
good dispersers) had the highest levels of mismatch (Fig. 5a, b).
Tree height showed the same pattern, with shorter species (rela-
tively poor dispersers) having the smallest levels of mismatch,
while the tallest species had the highest levels of mismatch
(Fig. 5a, b).
Mismatches for July mean temperature were generally posi-
tive, indicating that species latitudinal limits were warmer than
expected based on elevational limits. However, the mismatch
was significantly less for wind-dispersed and taller species
(Fig. 5c, d). A similar pattern was observed for mismatches in
FFD, with shorter species and those with unassisted seed disper-
sal having a greater number of FFD at their latitudinal than their
elevational limits (Fig. 5e, f). Mismatches were generally smaller
for taller species and those with animal-dispersed seed, but there
was a slight tendency for tall and wind-dispersed species to have
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 587
negative mismatches, indicating that these species had latitudi-
nal limits with fewer FFD than their elevational limits.
DISCUSSION
Comparison of species range limits along latitudinal and
elevational gradients of similar environmental change provides
a unique opportunity to study the factors controlling species
distributions. By comparing upper limits of tree species along a
latitudinal gradient in eastern North America and an elevational
gradient in the Great Smoky Mountains, our study investigates
the climatic factors that limit species ranges and how dispersal
traits may influence these limits. If the same climatic factors are
largely responsible for dictating range limits, there should be
close correspondence of species limits between gradients (Jump
et al., 2009; Randin et al., 2013; Kollas et al., 2014). Large
degrees of mismatch in climatic limits between the two gradi-
ents would suggest that factors other than, or in conjunction
with, climate were also responsible for controlling range limits
(McGlone, 1996; Ricklefs, 2004). We found a close correspond-
ence of climatic variables related to growing season length
and summer temperature at species upper latitudinal and
elevational range limits, consistent with the hypothesis that
common growing season length and temperature thresholds
limit species distributions along both gradients. In contrast, we
found large degrees of mismatch for variables associated with
winter temperature, suggesting that winter cold tolerance is not
the primary factor influencing range limits. Our results also
show that climatic mismatches between species latitudinal and
elevational range limits can be explained by traits related to
dispersal ability. The observed mismatch patterns are consistent
Figure 4 Relationship between climatic limits (2.5% quantile) of28 eastern North American tree species along latitudinal andelevational gradients for (a) January mean temperature, (b) Julymean temperature, and (c) frost free days. Solid lines representthe major axis regression line (regression coefficients given inTable 2). Species falling above the dotted 1:1 line have colderelevational limits than latitudinal limits. Species falling below thedotted line have warmer elevational limits than latitudinal limits.Species codes: Betula lenta (B.le.), Acer saccharum (A.sa.), Aesculusflava (A.fl.), Acer pensylvanicum (A. pe.), Prunus serotina (P. se.),Tsuga canadensis (T. ca.), Halesia tetraptera (H. te.), Acer rubrum(A. ru.), Quercus rubra (Q. ru.), Robinia pseudocacia (R. ps.),Fraxinus americana (F. am.), Tilia americana (T. am.), Magnoliaacuminate (M. ac.), Sassafras albidum (S. al.), Quercus montana(Q. mo.), Quercus velutina (Q. ve.), Oxydendrum arboretum(O. ar.), Nyssa sylvatica (N. sy.), Carya glabra (C. gl.), Carya alba(C. al.), Pinus rigida (P.ri.), Cornus florida (C. fl.), Liriodendrontulipifera (L. tu.), Quercus coccinea (Q. co.), Pinus strobus (P. st.),Pinus virginiana (P. vi.).
Table 2 Slope and R2 from major axis regression analysis ofelevational versus latitudinal climatic limits for 28 eastern NorthAmerican tree species. Climate variables are January meantemperature, July mean temperature, annual minimumtemperature, frost free days (FFD), growing season length (GSL),growing season temperature (GST) and growing degree days(GDD). The mean mismatch across the 28 species betweenlatitudinal and elevational limits is shown for each climatevariable. Values in parentheses are for the same analysis withoutlier (Pinus strobus) removed.
Variable Slope R2 Mismatch
Jan mean
temp (°C)
0.03* (0.08***) 0.08 (0.49) 7.57*** (7.76***)
Jul mean
temp (°C)
0.38** (0.52***) 0.18 (0.60) −0.97** (−0.97**)
Ann min
temp (°C)
0.06** (0.12***) 0.11 (0.51) 6.91*** (7.05***)
FFD 0.35** (0.52***) 0.18 (0.61) −0.41 (−0.38)
GSL 0.10* (0.26***) 0.08 (0.51) 59.3*** (62.1***)
GST 1.06** (0.92***) 0.20 (0.54) −1.37*** (−1.38***)
GDD 0.24** (0.42***) 0.14 (0.59) 129.6 (146.1*)
Significant at: *P < 0.1, **P < 0.05, ***P < 0.001.
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd588
with predictions under the hypothesis that historical dispersal
limitation restricts the poleward range limits of poorly dispers-
ing species.
Climatic limits
Climatic limitations on species range limits have the potential to
operate in several ways. First, low winter temperatures may limit
species ranges by causing freezing injury. For example, Sakai &
Weiser (1973) found that the freezing resistance of North
American trees correlated well with minimum winter tempera-
ture at their poleward range limits, providing evidence that cold
hardiness was a major factor determining species distributions.
In contrast, our finding of large mismatches in winter tempera-
tures between species latitudinal and elevational limits suggests
that cold tolerance was not the primary limiting factor govern-
ing range limits. In particular, winter temperatures were much
warmer at upper elevational limits compared with poleward
limits for most species, suggesting that species distributions in
the Great Smoky Mountains were not limited by winter tem-
peratures. Kollas et al. (2014) found similar results in a compari-
son of elevational and latitudinal gradients in Europe, where
there were also high degrees of mismatch between species limits
with regard to winter temperatures for seven broad-leaved tree
species. Our results do not rule out the possibility that cold
winter temperatures constrain species at high latitudes, either
alone or through interactions with other factors, but previous
studies of temperate tree species in Europe and North America
have found that poleward limits are better explained by other
climatic variables such as spring frost or growing season length
(Morin et al., 2007; Kollas et al., 2014).
Low summer or growing season temperatures have previously
been shown to constrain species distributions by limiting
growth and development (Pigott & Huntley, 1981; Mellert et al.,
2011). For example, Pigott & Huntley (1981) found that the
poleward limit of Tilia cordata (small-leaved basswood) in
Britain corresponded with summer temperatures that were too
low to permit fertilization and ovule development. Summer
temperatures have also been shown to explain the upper
elevational limits of tree species in the Bavarian Alps (Mellert
et al., 2011), elevations of spruce–fir ecotones of mountain
ranges in eastern North America (Cogbill & White, 1991) and
the position of alpine tree lines world-wide (Körner, 1998;
Körner & Paulsen, 2004). Short growing season length, alone or
Figure 5 Relationships between dispersaltraits (dispersal mode (a,c,e) andmaximum height (b,d,f)) and climaticmismatch between latitudinal andelevational range limits for 28 easternNorth American tree species. Results areshown for mismatch in January meantemperature (a,b), July mean temperature(c,d), and frost free days (e,f). Positivemismatches indicate colder elevationallimits than latitudinal limits. Negativemismatches indicate warmer elevationallimits than latitudinal limits. A zero value(dotted line) indicates no mismatch. Forpanels (a), (c) and (e) thick black barsrepresent median mismatch values, boxesrepresent first and third quartiles, whiskersextend to the most extreme values that arewithin 1.5 times the interquartile rangefrom the box and open circles representoutliers. For panels (b), (d) and (f) solidblack lines represent best fit lines fromsimple linear regression. Results of linearmodels (P and R2) are shown for eachtrait–mismatch relationship.
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 589
in combination with a low growing season temperature, may
also place physiological limitations on tree species. Using a
process-based model, Morin et al. (2007) determined that
poleward distributions of temperate trees in North America
were most often limited by insufficient GDD to undergo fruit
ripening or flowering. Similarly, Normand et al. (2009) found
that a short growing season limited the upper latitudinal and
elevational range limits of European plants. We found similar
results in our study, with July mean temperature and growing
season length limits having the highest correspondence between
the latitudinal and elevational gradients. These findings support
the hypothesis that common thresholds for growing season tem-
perature and/or length influence species limits along both gra-
dients. The close correlation between growing season length and
summer temperature in our study system makes it difficult to
disentangle their effects using our approach. Experimental tests
are required to determine whether growing season length,
summer temperature or a combination of the two most strongly
limit tree establishment and growth at species poleward and
high-elevation range limits.
Causes of mismatches between elevational andlatitudinal limits
Although summer temperature and growing season length gen-
erally corresponded well between species upper latitudinal and
elevational limits there was not a perfect match for all species.
We found that the degree of mismatch was strongly linked to
species dispersal traits, with poorer-dispersing species – those
with unassisted seed dispersal and low maximum height –
tending to have lower, and thus warmer, latitudinal limits than
expected based on their elevational climatic limits. This pattern
is consistent with the hypothesis that historical dispersal limita-
tion may have influenced the poleward distributions of species
with limited dispersal ability. The species with mismatch pat-
terns most consistent with the historical dispersal limitation
hypothesis include Aesculus flava, Halesia tetraptera, Robinia
pseudoacacia and Oxydendrum arboreum. These species had lati-
tudinal limits well below their elevational climatic thresholds
and traits associated with poor dispersal ability (short to
medium height and unassisted dispersal or animal-dispersed
seed). Further, their ancestral distributions are located in the
southern Appalachians (Burns & Honkala, 1990), meaning that
they have had far more time to disperse up the elevational gra-
dient in the Great Smoky Mountains than to move poleward
along the latitudinal gradient following the last glacial period.
Post-glacial migration lag has been implicated as a leading cause
of European tree species not filling their potential climatic
niches (Svenning & Skov, 2004) and has also been proposed to
limit the poleward distribution of Picea glauca (white spruce) in
eastern North America (Payette, 2007).
An alternative explanation for mismatches between species
elevational and latitudinal climatic limits is that long dispersal
distances relative to the scale of climatic variation in mountain-
ous regions allow species to exceed their fundamental climatic
niches through source–sink dynamics (Pulliam, 1988; Halbritter
et al., 2013). We cannot completely rule out this possibility in
our study, but we do not believe it is a major cause of the
observed mismatches. If species were able to extend their
elevational limits upwards by maintaining sink populations
above their long-term climatic limits, we would expect mis-
matches between elevational and latitudinal limits to be higher
for species with the greater dispersal ability, the opposite of the
observed pattern. Another possible explanation for why species
with poor dispersal ability had warmer latitudinal than
elevational limits is that low gene flow along the broad latitudi-
nal gradient may restrict genetic variation, and thus adaptation,
in range edge populations (Alleaume-Benharira et al., 2006). We
believe that this is an unlikely cause of the observed mismatch
patterns, for two reasons. First, population genetics studies show
that gene flow is extensive for most forest tree species
(Savolainen et al., 2007). Second, if low gene flow limits
poleward range expansion we would expect mismatches to be
greatest for species with low pollen dispersal ability, since gene
flow by pollen dispersal is typically much more extensive than
gene flow by seed dispersal (Ennos, 1994). However, we
observed no relationship between species pollen dispersal mode
and climatic mismatches.
Interestingly, one species, Pinus strobus (white pine), showed a
significantly different pattern from any of our other study
species, with an elevational limit well below its latitudinal cli-
matic thresholds. Pinus strobus can be categorized as a good
disperser based on its height and dispersal mode (Table S1), and
its observed mismatch pattern is consistent with the hypothesis
that high dispersal rates, and hence high gene flow, along the
short elevational gradient prevent adaptation to cold tempera-
tures found at high elevations. Previous studies have found
limited genetic differentiation among populations along
elevational gradients (Ohsawa & Ide, 2008), perhaps due to
extensive pollen flow. For example, Sundblad & Andersson
(1995) found no differentiation in cold hardiness in Pinus
sylvestris (Scots pine) across an elevational gradient, despite sig-
nificant differentiation across the latitudinal range of the
species. Although no studies to our knowledge have compared
the cold hardiness of P. strobus populations across elevational or
latitudinal gradients, common garden studies have shown that
southern Appalachian seed sources grow faster than individuals
from any other part of P. strobus’s range (Wendel & Smith,
1990). The failure of P. strobus to grow at high elevations in the
Great Smoky Mountains may reflect the tradeoff between cold
hardiness and growth observed in many tree species (Loehle,
1998; Koehler et al., 2012), such that fast-growing, low-elevation
genotypes are not capable of surviving at higher elevations with
colder temperatures and a shorter growing season. Further tests
of this hypothesis would involve direct comparisons of gene
flow and local adaptation to climate in P. strobus populations
along the latitudinal and elevational gradients.
CONCLUSIONS
Our results demonstrate that there is a positive relationship
between the latitudinal and elevational climatic limits of eastern
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd590
North American trees – species that grow at higher latitudes also
grow at higher elevations. Comparison of the high-latitude and
high-elevation limits of tree species suggests that low summer
temperatures and a short growing season length limit species
distributions along both gradients, whereas winter temperature
probably does not play a critical role. We also demonstrate that
the degree of climatic mismatch between elevational and latitu-
dinal range limits can be explained by species dispersal traits. In
particular, species with poor dispersal ability did not grow as far
poleward as expected based on their elevational limits, suggest-
ing that historical dispersal constraints may have limited the
post-glacial expansion of some species. These findings have
implications for how individual species may respond to current
and ongoing climate shifts. Based on climate predictions for the
coming century, all of the climatic range limits we identify here
are expected to shift poleward and up in elevation (Colwell et al.,
2008). Our analysis suggests that poorly dispersing species may
not adequately track climate shifts along the latitudinal gradient
in eastern North America. The inability to keep pace with suit-
able climate space combined with the contraction of trailing
range edges may put these species at increased risk compared
with better dispersers.
ACKNOWLEDGEMENTS
We are grateful to Alaä Craddock and Keith Langdon for their
help compiling the GSMNP temperature dataset. We thank
Jannice Friedman and Alex Twyford for stimulating discussions
on gene flow and local adaptation and Alistair Jump and an
anonymous reviewer for providing constructive comments on
an earlier version of the manuscript. A.S. was supported by
a National Science Foundation Graduate Research Fellowship.
REFERENCES
Alleaume-Benharira, M., Pen, I.R. & Ronce, O. (2006) Geo-
graphical patterns of adaptation within a species’ range: inter-
actions between drift and gene flow. Journal of Evolutionary
Biology, 19, 203–215.
Billings, W. (1973) Arctic and alpine vegetations: similarities,
differences, and susceptibility to disturbance. BioScience, 23,
697–704.
Burns, R.M. & Honkala, B.H. (1990) Silvics of North America.
Agricultural Handbook 654. US Department of Agriculture
Forest Service, Washington, DC.
Busing, R.T., Stephens, L.A. & Clebsch, E.E.C. (2005) Climate
data by elevation in the Great Smoky Mountains: a database
and graphical displays for 1947–1950 with comparison to
long-term data. Data Series Report DS 115. US Geological
Survey, Washington, DC.
Chen, I.-C., Hill, J.K., Ohlemüller, R., Roy, D.B. & Thomas, C.D.
(2011) Rapid range shifts of species associated with high levels
of climate warming. Science, 333, 1024–1026.
Cogbill, C. & White, P. (1991) The latitude–elevation relation-
ship for spruce–fir forest and treeline along the Appalachian
mountain chain. Vegetatio, 94, 153–175.
Colwell, R.K., Brehm, G., Cardelús, C.L., Gilman, A.C. &
Longino, J.T. (2008) Global warming, elevational range shifts,
and lowland biotic attrition in the wet tropics. Science, 322,
258–261.
Davis, M., Woods, K., Webb, S. & Futyma, R. (1986) Dispersal
versus climate: expansion of Fagus and Tsuga into the Upper
Great Lakes region. Vegetatio, 67, 93–103.
Ennos, R. (1994) Estimating the relative rates of pollen and
seed migration among plant populations. Heredity, 72, 250–
259.
Fridley, J.D. (2009) Downscaling climate over complex terrain:
high finescale (<1000 m) spatial variation of near-ground
temperatures in a montane forested landscape (Great Smoky
Mountains). Journal of Applied Meteorology and Climatology,
48, 1033–1049.
Fridley, J.D. (2010) An inventory and archive of historical tem-
perature and precipitation data for Great Smoky Mountains
National Park. Technical Report. US National Park Service,
Washington, DC.
García-Ramos, G. & Kirkpatrick, M. (1997) Genetic models of
adaptation and gene flow in peripheral populations. Evolu-
tion, 51, 21–28.
Halbritter, A.H., Alexander, J.M., Edwards, P.J. & Billeter, R.
(2013) How comparable are species distributions along
elevational and latitudinal climate gradients? Global Ecology
and Biogeography, 22, 1228–1237.
Higgens, S.I., Nathan, R. & Cain, M.L. (2003) Are long-distance
dispersal events in plants usually caused by nonstandard
means of dispersal? Ecology, 84, 1945–1956.
Horvitz, C.C. & Le Corff, J. (1993) Spatial scale and dispersion
pattern of ant- and bird-dispersed herbs in two tropical
lowland rain forests. Vegetatio, 107, 351–362.
von Humboldt, A. (2009) Essay on the geography of plants (ed. by
S.T. Jackson and transl. by S. Romanowski). University of
Chicago Press, Chicago, IL.
Huntley, B., Bartlein, P.J. & Prentice, I.C. (1989) Climatic
control of the distribution and abundance of beech (Fagus L.)
in Europe and North America. Journal of Biogeography, 16,
551–560.
Jobbágy, E.G. & Jackson, R.B. (2000) Global controls of forest
line elevation in the Northern and Southern Hemispheres.
Global Ecology and Biogeography, 9, 253–268.
Johnson, W. & Webb, T.I. (1989) The role of blue jays
(Cyanocitta cristata L.) in the postglacial dispersal of
fagaceous trees in eastern North America. Journal of Biogeog-
raphy, 16, 561–571.
Jump, A.S., Mátyás, C. & Peñuelas, J. (2009) The altitude-for-
latitude disparity in the range retractions of woody species.
Trends in Ecology and Evolution, 24, 694–701.
Koehler, K., Center, A. & Cavender-Bares, J. (2012) Evidence for
a freezing tolerance–growth rate trade-off in the live oaks
(Quercus series Virentes) across the tropical–temperate divide.
New Phytologist, 193, 730–744.
Kollas, C., Körner, C. & Randin, C.F. (2014) Spring frost and
growing season length co-control the cold range limits of
broad-leaved trees. Journal of Biogeography, 41, 773–783.
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 591
Körner, C. (1998) A re-assessment of high elevation treeline
positions and their explanation. Oecologia, 115, 445–459.
Körner, C. & Paulsen, J. (2004) A world-wide study of high
altitude treeline temperatures. Journal of Biogeography, 31,
713–732.
Little, E.L. (1971) Atlas of United States trees. Miscellaneous Pub-
lication 1146. US Department of Agriculture Forest Service,
Washington, DC.
Loehle, C. (1998) Height growth rate tradeoffs determine north-
ern and southern range limits for trees. Journal of Biogeogra-
phy, 25, 735–742.
McGlone, M.S. (1996) When history matters: scale, time, and
tree climate diversity. Global Ecology and Biogeography Letters,
5, 309–314.
MacKenzie, M. & White, P. (1998) Vegetation of Great Smoky
Mountains National Park, 1935–1938. Castanea, 63, 323–336.
Meehl, G.A., Arblaster, J.M. & Branstator, G. (2012) Mecha-
nisms contributing to the warming hole and the consequent
U.S. east–west differential of heat extremes. Journal of Climate,
25, 6394–6408.
Mellert, K.H., Fensterer, V., Küchenhoff, H., Reger, B., Kölling,
C., Klemmt, H.J. & Ewald, J. (2011) Hypothesis-driven species
distribution models for tree species in the Bavarian Alps.
Journal of Vegetation Science, 22, 635–646.
Morin, X., Augspurger, C. & Chuine, I. (2007) Process-based
modeling of species’ distributions: what limits temperate tree
species’ range boundaries? Ecology, 88, 2280–2291.
Nathan, R. (2006) Long-distance dispersal of plants. Science,
313, 786–788.
Nathan, R., Katul, G.G., Horn, H.S., Thomas, S.M., Oren, R.,
Avissar, R., Pacala, S.W. & Levin, S.A. (2002) Mechanisms of
long-distance dispersal of seeds by wind. Nature, 418, 409–
413.
Nathan, R., Schurr, F.M., Spiegel, O., Steinitz, O., Trakhtenbrot,
A. & Tsoar, A. (2008) Mechanisms of long-distance
seed dispersal. Trends in Ecology and Evolution, 23, 638–
647.
Normand, S., Treier, U.A., Randin, C., Vittoz, P., Guisan, A. &
Svenning, J.-C. (2009) Importance of abiotic stress as a range-
limit determinant for European plants: insights from species
responses to climatic gradients. Global Ecology and Biogeogra-
phy, 18, 437–449.
Ohsawa, T. & Ide, Y. (2008) Global patterns of genetic variation
in plant species along vertical and horizontal gradients on
mountains. Global Ecology and Biogeography, 17, 152–163.
Payette, S. (2007) Contrasted dynamics of northern Labrador
tree lines caused by climate change and migrational lag.
Ecology, 88, 770–780.
Pigott, C. & Huntley, J. (1981) Factors controlling the distribu-
tion of Tilia cordata at the northern limits of its geographical
range. III Nature and cause of seed sterility. New Phytologist,
87, 817–839.
Portmann, R.W., Solomon, S. & Hegerl, G.C. (2009) Spatial and
seasonal patterns in climate change, temperatures, and pre-
cipitation across the United States. Proceedings of the National
Academy of Sciences USA, 106, 7324–7329.
Pulliam, H.R. (1988) Sources, sinks, and population regulation.
The American Naturalist, 132, 652–661.
Randin, C.F., Paulsen, J., Vitasse, Y., Kollas, C., Wohlgemuth, T.,
Zimmermann, N.E. & Körner, C. (2013) Do the elevational
limits of deciduous tree species match their thermal latitudi-
nal limits? Global Ecology and Biogeography, 22, 913–923.
Ricklefs, R.E. (2004) A comprehensive framework for global
patterns in biodiversity. Ecology Letters, 7, 1–15.
Royal Botanic Gardens Kew (2014) Seed information database
(SID), Version 7.1. Available at: http://data.kew.org/sid/
(accessed 1 March 2014).
Sakai, A. & Weiser, C. (1973) Freezing resistance of trees in
North America with reference to tree regions. Ecology, 54,
118–126.
Savolainen, O., Pyhäjärvi, T. & Knürr, T. (2007) Gene flow and
local adaptation in trees. Annual Review of Ecology, Evolution,
and Systematics, 38, 595–619.
Schimper, A.F.W. (1898) Plant geography upon a physiological
basis. Clarendon Press, Oxford.
Shanks, R.E. (1954) Climates of the Great Smoky Mountains.
Ecology, 35, 354–361.
Sundblad, L. & Andersson, B. (1995) No difference in frost har-
diness between high and low altitude Pinus sylvestris (L.)
offspring. Scandinavian Journal of Forest Research, 10, 22–
26.
Svenning, J., Normand, S. & Skov, F. (2008) Postglacial dispersal
limitation of widespread forest plant species in nemoral
Europe. Ecography, 31, 316–326.
Svenning, J.-C. & Skov, F. (2004) Limited filling of the potential
range in European tree species. Ecology Letters, 7, 565–573.
Thomson, F.J., Moles, A.T., Auld, T.D. & Kingsford, R.T. (2011)
Seed dispersal distance is more strongly correlated with plant
height than with seed mass. Journal of Ecology, 99, 1299–1307.
Tranquillini, W. (1979) Physiological ecology of the alpine timber-
line. Tree existence at high altitudes with special references to the
European Alps. Springer, Berlin.
Wendel, G.W. & Smith, H.C. (1990) Pinus strobus L. Silvics of
North America (ed. by R.M. Burns and B.H. Honkala), pp.
476–488. Agricultural Handbook 654. US Department of
Agriculture Forest Service, Washington, DC.
Whittaker, R. (1956) Vegetation of the Great Smoky Mountains.
Ecological Monographs, 26, 1–80.
Wilson, M.F. & Traveset, A. (2000) The ecology of seed dispersal.
Seeds: the ecology of regeneration in plant communities (ed. by
M. Fenner), pp. 85–110. CABI Publishing, New York.
SUPPORTING INFORMATION
Additional supporting information may be found in the online
version of this article at the publisher’s web-site.
Figure S1 Climate trends (1900–2011) in Great Smoky Moun-
tain National Park for January mean temperature, July mean
temperature, and frost free days.
Figure S2 Distributions of study species in eastern North
America and along the elevational gradient in the Great Smoky
Mountains.
A. Siefert et al.
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd592
Table S1 Dispersal traits of study species.
Table S2 P-values for effects of species traits on latitude–
elevation climatic mismatch.
BIOSKETCHES
Andrew Siefert is a post-doctoral researcher at the
University of California, Davis. He has interests in plant
community assembly and functional ecology.
Mark Lesser is a post-doctoral researcher at the State
University of New York College of Environmental
Science and Forestry. His research interests are in
population and community ecology.
Jason Fridley is a plant ecologist with interests in
community assembly and ecosystem functioning from
local to global scales.
Editor: Joesp Penuelas
Latitudinal and elevational range limits
Global Ecology and Biogeography, 24, 581–593, © 2015 John Wiley & Sons Ltd 593